Impacts of combustion phasing, load, and speed on soot formation in spark-ignition engines

The premixed prevaporized engine operation method was used to study the effects of main combustion thermodynamic properties and residence time on soot formation in a spark-ignition engine. Select cases were repeated under early-injection, nearly homogeneous, spark-ignition direct-injection operation to determine if the impact of the investigated parameters was the same or if the impact of in-cylinder liquid fuel injection and the resulting heterogeneous fuel-air mixture alters the trends. The original premixed prevaporized study hypothesized that soot is more likely to form after main combustion than during the main combustion event under completely homogeneous conditions. This hypothesis was tested in this study by performing premixed prevaporized combustion phasing sweeps at equivalence ratios (phi s) of 1.35 and 1.40. Both sweeps showed low sensitivity of the particle size distribution to significant changes in peak temperature and pressure during combustion, providing supporting evidence for the original hypothesis. This information was then used to design experiments to isolate the impacts of pressure (engine load) and residence time (engine speed). A premixed prevaporized load sweep showed that particulate emissions increase as a function of load/pressure. A spark-ignition direct-injection load variation showed similar pressure dependence for cases with in-homogeneous in-cylinder fuel-air distributions. A premixed prevaporized residence time variation (performed by changing engine speed) demonstrated an increase in soot formation with increased residence time. The results for identical spark-ignition direct-injection residence-time variations suggest a trade-off in soot formation between the effects of increased mixing time and increased residence time for spark-ignition direct-injection operation. The premixed prevaporized load and speed points were each investigated using phi sweeps to determine the critical enrichment threshold for soot formation (phi(C)) and the dependence of soot formation for phi > phi(C). The spark-ignition direct-injection investigations were performed at phi = 0.98, such that any soot formation above the non-fuel-related baseline particle size distribution could be attributed either to mixture heterogeneity or in-cylinder fuel films.